由於末期癌症的死亡率相當高，所以需要能在初期就將癌症給診斷出來的技術及方法，近年來在質譜技術上的進步使得該診斷的可能性大大的增加，因此高解析度質譜圖資料的分析逐漸被應用在疾病分類及診斷指導。質譜儀器在解析度及校正上的誤差使得質譜圖資料產生質量峰偏移的現像(peak shifting)，而一般來說質譜圖上的質量峰(peak)又高達幾十萬個，這兩個問題使得分析質譜圖時產生困難而且要花費大量的時間，我們提出以區間最大值的方式來處理質量峰偏移現像並利用統計分析快速的在幾十萬個質量峰中找到較少量且具辨別力的質量峰(biomarker candidates)。最後應用所提出之方法在長庚大學蛋白質體質心實驗室所提供的口腔癌上做測試。 Due to the high death rate in advanced stage diseases, the diagnosis of early-stage cancer is needed for public health. Recent advances in the biotechnology of high-throughput and high-resolution MALDI-TOF mass spectrometry (MS) has made such diagnosis possible (Petricoin and Liotta 2003). From then on, high-resolution mass spectrometers are increasingly used for disease classification and therapeutic guidance. Due to instrument resolution and/or instrument calibration, the mass spectrometry (MS) data may be poor in quality. The problem makes it difficult and time consuming to trace each spectrum with thousands of features for biomarkers. We proposed a region-based alignment method to deal with peak shifting problem and applied statistical method to select most discriminatory peaks as biomarker candidates. Finally, we test our methodology on the oral cancer dataset from CHANG GUNG UNIVERSITY in Taiwan.
